Science (2013) 20, 228–234, DOI 10.1111/j.1744-7917.2011.01495.x

ORIGINAL ARTICLE Pollinating fig wasp solmsi adjusts the offspring sex ratio to other foundresses

Hao-Yuan Hu1, Zhong-Zheng Chen1, Zi-Feng Jiang2, Da-Wei Huang3,4, Li-Ming Niu5 and Yue-Guan Fu5 1Key Laboratory of Biotic Environment and Ecological Safety in Anhui Province, College of Life Sciences, Anhui Normal University, Wuhu, Anhui Province, China, 2Department of Biology, University of Maryland, College Park, Maryland, USA, 3Key Laboratory of Zoological Systematics and Evolution, Institute of Zoology, Chinese Academy of Sciences, Beijing, China, 4Plant Protection College, Shandong Agricultural University, Tai’an, Shandong Province, China, 5Environment and Plant Protection Institute, Chinese Academy of Tropical Agricultural Sciences, Danzhou, Hainan Province, China

Abstract Local mate competition theory predicts that offspring sex ratio in pollinating fig wasps is female-biased when there is only one foundress, and increased foundress density results in increased offspring sex ratio. Information of other foundresses and clutch size have been suggested to be the main proximate explanations for sex ratio adjustment under local mate competition. Our focus was to show the mechanism of sex ratio adjustment in a pollinating , Mayr, an obligate pollinator of the functionally dioecious fig, hispida Linn., with controlled experiments in the field. First, we obtained offspring from one pollinator and offspring at different oviposition sequences, and found that offspring sex ratio decreased with clutch size, and pollinators produced most of their male offspring at the start of bouts, followed by mostly females. Second, we found that offspring sex ratio increased with foundress density, and pollinators did adjust their offspring sex ratio to other females in the oviposition patches. We suggest that when oviposition sites are not limited, pollinators will mainly adjust their offspring sex ratio to other foundresses independent of clutch size changes, whereas adjusting clutch size may be used to adjust sex ratio when oviposition sites are limited. Key words behavior, co-evolution, local mate competition, mutualism, pollinating fig wasp, sex ratio adjustment

Introduction highly structured population, such as mating only between full siblings, evolutionary stable strategy models predict a Sex ratio in a local environment has been taken as a model female-biased sex ratio (Hamilton, 1967; Maynard Smith, of sex ratio theory (Hamilton, 1967; Frank, 1985; Herre, 1976). In haplodiploid , with haploid males 1985, 1987; West et al., 2000; Steiner & Ruther, 2009; (developed from unfertilized eggs) and diploid females West, 2009) and a case study of adaptation in evolution- (developed from fertilized eggs), the primary sex ratio is ary biology (Fellowes et al., 1999). In a random-mating determined by the mothers’ control of fertilization at the populations of infinite size, natural selection favors equal time of oviposition (Charnov, 1982). investment in both sexes, and an equal sex ratio is an There may be natural selection on foundresses to adjust evolutionary stable result (Fisher, 1930). However, in a their offspring sex ratio because of local mate competition (LMC) (Charnov, 1982; West et al., 2005). When patches are founded by one female and all mating occurs between Correspondence: Hao-Yuan Hu, College of Life Sciences, siblings, brothers may compete locally with each other Anhui Normal University, Wuhu, Anhui Province 241000, for mating with their sisters. Such LMC selects for fewer China. email: [email protected] male offspring (Sheldon & West, 2002). If offspring in a

C 2012 The Authors 228 Insect Science C 2012 Institute of Zoology, Chinese Academy of Sciences Sex ratio adjustment of a fig wasp 229 patch originate from two foundresses, the value of sons Academy of Tropical Agricultural Sciences (CATAS), increases because sons have the potential to mate with the Danzhou, Hainan province, China (19◦30.410S, daughters of other foundresses, and a less female-biased 109◦29.340E). Hainan Island is located south of the offspring sex ratio is preferred (Hamilton, 1967, 1979; China mainland separated by the Qiongzhou Strait, ap- Herre, 1985, 1987; Werren, 1987). Sex ratios increase proximately 40 km wide. The climate includes well- with foundress number, as LMC models predict, in many defined dry (November–April) and rainy (May–October) parasitoid wasps (reviewed in King, 1993) and some fig seasons, and the annual mean temperature is 24.3◦C, with wasps (e.g., Raja et al., 2008) the lowest mean temperature in February (18.2◦C) and the The proximate explanations for such sex ratio ad- highest in July (29.6◦C). justment under LMC may differ among species (Moore is a functionally dioecious and free- et al., 2005). Some reports have argued that clutch size standing tree, with inter-sexually synchronously flowering might be one of the main explanations when more par- phenologies. Most fruits grow on fruit branches, and can asitoid wasps or fig wasps oviposit in a limited patch be kept away from pollinators using specially made se- (Kjellberg et al., 2005; Moore et al., 2005; Raja et al., curing nylon bags during fig receptivity. Pollinators can 2008). When multiple foundresses oviposit in figs, ovipo- be collected from mature male fruits in mesh-lidded pots sition site limitation would lead to reduction of offspring just before wasp emergence (Hu et al., 2009, 2010). reproduction (Moore & Greeff, 2003). Limited clutch size and the oviposition order of fertilized and unfertilized eggs may lead to a less female-biased sex ratio (Kjellberg Offspring sex ratio and clutch size et al., 2005; Moore et al., 2005; Raja et al., 2008; Sun et al., 2009). Another explanation of sex ratio adjustment To investigate the effect of clutch size on offspring sex under LMC is foundress density affecting probability of ratio, we obtained offspring from one-foundress figs. Fe- ovipositing a son and that probability being independent male pollinators were collected as above and immedi- of order. Foundresses may adjust their offspring sex ratios ately introduced into the protected receptive figs with a by the information of other foundresses or of oviposi- soft brush. One pollinator was allowed to enter each fig, tion in a patch left by other foundresses (Hamilton, 1967, and then the figs were re-bagged individually; 241 one- 1979; Herre, 1985, 1987; Kinoshita et al., 2002; Moore foundress figs were obtained from seven male F. hispida et al., 2002; Pereira & Prado, 2006; Herre et al., 2008; tress from January to December in 2007. Abe et al., 2009; Somjee et al., 2011). Small clutch size can be obtained by ceasing oviposi- In some pollinating fig wasps, clutch size would be tion after the pollinator’s entry. Pollinators oviposit inside the major explanation to sex ratio adjustment under LMC figs, and it is hard to cease the oviposition and keep devel- (Kjellberg et al., 2005; Moore et al., 2005; Raja et al., opment of the figs at the same time. Recent research has 2008; Sun et al., 2009). The effect of sex adjustment to used poisons such as insecticide pyrethrum (Raja et al., other foundresses has been called in to question (Kjellberg 2008) and ether (Sun et al., 2009) to cease the oviposition. et al., 2005). But oviposition sites would not be always Patino et al. (1994) showed that when figs reach temper- limited when only several foundresses lay eggs in a patch. atures higher than 8◦C above the ambient temperature When oviposition sites in a patch are sufficient, clutch size in full sunlight, pollinators inside figs would be killed. should be unaffected by foundress density. Given this, we Warm water has been used to kill pollinators inside figs predict that the effect of clutch size on sex ratio adjustment (Hu et al., 2009). In this article, warm water (36–40◦C), should be weak when oviposition sites are rarely limited, 8◦C above ambient temperature, was injected into a plas- as in some figs with plenty of florets, where oviposition tic bag (≈ 2 L), and the figs were put into the bag to sites are always enough. We test this with Ceratosolen kill pollinators in situ. One-foundress figs were opened at solmsi Mayr, an obligate pollinator of the functionally 2-min intervals after the figs were put into warm water to dioecious fig, Ficus hispida Linn. examine whether the foundress was dead or not. Ten min- utes were selected because foundresses inside large and small figs were all dead then. Twenty one-foundress figs Materials and methods warmed at 4 h after pollinator entry were protected from entry by other foundresses with nylon bags, and 13 treated Study species and site figs matured normally. Previous research has shown that C. solmsi could only live in figs for less than 24 h, and Experiments on Ficus hispida Linn. – Ceratosolen the duration of oviposition lasts less than 9 h (Yang et al., solmsi Mayr mutualism were conducted at the Chinese 2002). Thus, three treatments were designed: one without

C 2012 The Authors Insect Science C 2012 Institute of Zoology, Chinese Academy of Sciences, 20, 228–234 230 H. Y. Hu et al. warming, the other two with 2- and 4-h intervals (from pollinator entering to warming) respectively. More than 50 one-foundress figs were used in each treatment on two male F. hispida trees in September 2007 and August 2008, respectively.

Offspring sex ratio and foundress density

To investigate the effect of foundress density on the offspring sex ratio, the following experiments were conducted. Five treatments were designed with 1–5 foundresses respectively, and more than 30 figs were used in each treatment. Female pollinators were collected as above and immediately introduced into the protected re- ceptive figs. Wasps introduced to multiple-foundress figs Fig. 1 Correlation between clutch size and sex ratio of Cer- were from mature figs on different fig trees, which en- atosolen solmsi in one-foundress figs. Offspring sex ratio de- sured that the wasps were non-siblings. In the multiple- creased with clutch size (F1,230 = 13.45, P < 0.001, r =−0.24). foundress figs, foundresses were allowed to enter each fig at 20-min intervals so the earlier entrants could finish crawling through most of the ostiolar bracts. The experi- The sex ratio data were arcsine square root transformed ments were conducted on five male F. h i s p i d a trees from and used as the response variable in the analyses assuming January to December in 2007. normal errors (Flanagan et al., 1998; Crawley, 2007). The Oviposition sites would not be limited in multiple- other analyses were also finished in R2.6.1. foundress figs if the entered foundresses were killed and oviposition inside figs ceased. One-foundress and Results three-foundress figs were obtained. Two hours after the first foundresses were introduced into figs, the figs were Clutch size and offspring sex ratio in figs with one warmed and oviposition insides figs ceased with the meth- foundress ods as above. More than 30 figs were used in each treat- ment on a male F. hispida tree in August 2008. Of the 241 one-foundress figs, nine of them had only male offspring and were removed for further analy- Data analyses ses. Mean numbers of female florets and galls (includ- ing empty galls) per fig were 1 766.19 ± 300.66 and When the figs matured, female and male wasps, galls, 312.96 ± 73.45, respectively. Mean numbers of male and total female flowers in each fig were counted. To- and female offspring per fig were 31.14 ± 16.34 and tal female flowers were counted from a quarter part of 194.56 ± 89.47, respectively. Thus, female florets were the fig. The proportional data, such as sex ratio, often more than one foundress could use. Offspring sex ratios have non-normally distributed errors and are difficult to were highly female-biased and mean sex ratio (male/total analyze with traditional analysis of variance (ANOVA) offspring) was 0.15 ± 0.07. Offspring sex ratio (arc- techniques (Hardy, 2002). To avoid these problems, we sine square root transformed) decreased with clutch size first analyzed the sex ratio data using generalized lin- (F1,230 = 13.45, P < 0.001, r =−0.24) (Fig. 1), which is ear model (GLM) analysis of deviance, assuming bino- consistent with the theoretical expectation that males are mial errors, and a logit link function in R2.6.1 (Ihaka & produced first. Gentleman, 1996; Crawley, 2007). The number of male To investigate the sex ratios at different oviposition se- wasps in a fig was used as the response variable, and the quences, we compared the offspring sex ratio at three total offspring number in a fig was used as the binomial time points: 2 h, 4 h and normal (life-long), and found denominator. We calculated a heterogeneity factor (HF), that sex ratios (arcsine square root transformed) were the ratio of residual deviance to the residual degrees of not significantly different (F2,102 = 1.46, P = 0.24). freedom, to examine the data for overdispersion. However, The number of male and female offspring both sig- HF in our analyses were all larger than 3, which means nificantly changed among treatments (F2,102 = 15.76, the data would not fit the assumption of binomial errors. P < 0.001; F2,102 = 38.43, P < 0.001, respectively).

C 2012 The Authors Insect Science C 2012 Institute of Zoology, Chinese Academy of Sciences, 20, 228–234 Sex ratio adjustment of a fig wasp 231

Fig. 2 Offspring sex ratio (square), male (triangle) and female Fig. 3 Offspring sex ratio of Ceratosolen solmsi and gall num- (circle) Ceratosolen solmsi offspring (means ± SD) in one- ber at different foundress densities. Offspring sex ratio (square) = foundress figs with 2- and 4-h oviposition duration. Sex ratio significantly increased with foundress density (F1,210 78.44, < = was not significantly different (F2,102 = 1.46, P = 0.24), and P 0.001, r 0.52). pollinators tended to produce most male offspring at the start of bouts, followed by mostly females. one-foundress figs with 2-h oviposition, sex ratio (arc- sine square root transformed) in three-foundress figs was significantly larger (t . = 2.55, P = 0.02) (Fig. 4). Male offspring significantly differed between 2-h and 4-h 19 89 treatments (least significant difference test [LSD], P < 0.001) but were similar between the 4-h treatment Discussion and normal (LSD, P = 0.61), and female offspring signif- icantly differed between treatments (LSD, all P < 0.001). Previous reports have shown that a female-biased sex Pollinators tended to produce most of their male offspring ratio is expected when there is one foundress oviposit- at the start of bouts, followed by mostly females (Fig. 2). ing in the fig cavity, and a lower female-biased sex ra- tio with increasing foundress density (Hamilton, 1967, Foundress density and offspring sex ratio 1979; Charnov, 1982; Herre, 1985, 1987; Werren, 1987). Our results show that offspring sex ratio significantly in- In the one-foundress and multi-foundress figs, floret creases with foundress density, and sex ratio adjustment numbers are similar (F4,207 = 1.54, P = 0.19). Offspring of the pollinator is in line with the theoretical prediction sex ratio (arcsine square root transformed) was signifi- of LMC models. Clutch size has been suggested to be one cantly different among the five treatments (F4,207 = 22.70, of the proximate explanations for sex ratio adjustment un- P < 0.001), and significantly increased with foundress der LMC (Kjellberg et al., 2005; Moore et al., 2005; Raja density (F1,210 = 78.44, P < 0.001, r = 0.52) (Fig. 3). et al., 2008; Sun et al., 2009). Some pollinating fig wasps Offspring sex ratio in figs with one or more foundresses produce most male offspring at the beginning of ovipo- with similar total numbers of offspring in the fig were sition (Moore et al., 2005; Kjellberg et al., 2005; Raja compared (Table 1). One-foundress figs usually had less et al., 2008; Sun et al., 2009). If multiple foundresses than 400 offspring and only three multi-foundress figs oviposit in a fig, there should be a reduction in clutch had less than 150. Therefore, only figs with between size per foundress because of limited oviposition sites, 151–400 offspring were analyzed. Sex ratios (arcsine and the LMC sex ratio increasing proportionately with square root transformed) in multi-foundress figs were all foundress density would be expected (Kjellberg et al., significantly higher than that in one-foundress figs when 2005; Moore et al., 2005; Raja et al., 2008; Sun et al., the number of total offspring in the two types of figs 2009). In the present study, a significantly negative rela- were similar (Table 1). tionship between offspring sex ratio and clutch size was In those three-foundress figs with 2-h oviposition dura- found in normal one-foundress figs. Pollinators tended to tion, mean numbers of galls, and male and female off- produce most of their male offspring at the start of bouts, spring per fig were 259.92 ± 61.16, 29.83 ± 11.82, followed by mostly females, which is similar to Lipor- 94.75 ± 65.95 (all n = 12), respectively. Compared to rhopalum tentacularis (Raja et al., 2008). In F. h i s p i d a

C 2012 The Authors Insect Science C 2012 Institute of Zoology, Chinese Academy of Sciences, 20, 228–234 232 H. Y. Hu et al.

Table 1 Comparisons of offspring sex ratio (means ± SD) of pollinators between figs with one or more foundresses with similar clutch size.

Range of Figs with Figs with more than t-test for means t-test for sex ratio (arcsine clutch size one foundress one foundress of clutch size square root transformed)

∗∗ 151–200 0.16 ± 0.03 (20) 0.24 ± 0.07 (11) t17.23 = 0.74, NS t13.67 = 3.73 ∗ 201–250 0.16 ± 0.04 (10) 0.26 ± 0.14 (13) t19.28 = 0.41, NS t15.69 = 2.67 ∗ 251–300 0.13 ± 0.03 (19) 0.23 ± 0.09 (9) t18.21 = 0.46, NS t9.03 = 3.22 ∗∗ 301–350 0.15 ± 0.07 (7) 0.30 ± 0.11 (10) t14.47 = 0.35, NS t14.06 = 3.56 ∗ 351–400 0.13 ± 0.04 (9) 0.23 ± 0.09 (6) t13 = 2.07, NS t6.44 = 2.55

The number in parentheses means sample size. NS, not significant at P > 0.05. ∗P < 0.05, ∗∗P < 0.01.

gard to other foundresses. After clutches from different foundresses have been distinguished quantificationally, Greeff & Newman (2011) illustrated that the pollinating fig wasp Platyscapa awekei simultaneously uses its own clutch size as well as the other females in allocating sex, and especially, the wasp mostly adjusts its sex ratio in response to total foundresses and their clutch sizes. Our results show that those multi-foundress figs, with similar clutch size to one-foundress figs, had significantly higher offspring sex ratio, which suggests that foundress density significantly affected offspring sex ratio, even indepen- dent of clutch size. When oviposition only lasted for 2 h by experimental design, there was significantly higher sex ratio in three-foundress figs than in one-foundress figs, which indicates that the explanation to the sex ratio ad- Fig. 4 Offspring sex ratio of Ceratosolen solmsi in one- justment under LMC should mainly be in regard to other foundress and three-foundress figs with 2-h oviposition dura- foundresses. tion. Sex ratio was significantly larger in three-foundress figs (t19.89 = 2.55, P = 0.02). Acknowledgments figs, there are enough florets, and oviposition site limita- This project was supported by the National Natural Sci- tion would be weak when only a few foundresses oviposit ence Foundation of China (NSFC grant no. 30770302, in the figs, in which case clutch size changes are unlikely 30570970), partially supported by Program of Ministry to provide a proximate mechanism for increased sex ratios of Science and Technology of the Republic of China in response to other mothers. However, we suggest that (2006FY110500), and by the National Science Fund for changes in clutch size as a mechanism for sex ratio adjust- Fostering Talents in Basic Research (Special subjects in ment under LMC of the pollinator would work when far , NSFC-J0630964/J0109). more foundresses enter into the fig cavity, because then oviposition sites would be limited. Disclosure Alternatively, foundresses may adjust their offspring sex ratio in regard to other foundresses or in regard to The authors declare that they have no conflict of interest. oviposition in a patch left by other foundresses indepen- dently of order and clutch size (Hamilton, 1967, 1979; Herre, 1985, 1987; Werren, 1987; Kinoshita et al., 2002; References Abe et al., 2009; Somjee et al., 2011). Given another foundress in a short interval (0.5 h), Kinoshita et al. Abe, J., Kamimura, Y., Shimada, M. and West, S.A. (2009) Ex- (2002) found contacting another foundress would cause tremely female-biased primary sex ratio and precisely con- a less female-biased sex ratio, which indicates pollinat- stant male production in a parasitoid wasp Melittobia. Animal ing fig wasps could adjust the offspring sex ratio in re- Behaviour, 78, 515–523.

C 2012 The Authors Insect Science C 2012 Institute of Zoology, Chinese Academy of Sciences, 20, 228–234 Sex ratio adjustment of a fig wasp 233

Charnov, E.L. (1982) The Theory of Sex Allocation. Princeton Kinoshita, M., Kasuya, E. and Yahara, T. (2002) Effects of time- University Press, Princeton, USA. 355 pp. dependent competition for oviposition sites on clutch sizes Crawley, M.J. (2007) The R Book. John Wiley & Sons Ltd., andoffspringsexratiosinafigwasp.Oikos, 96, 31–35. Chichester, UK. 950 pp. Kjellberg, F., Bronstein, J.L., van Ginkel, G., Greeff, Fellowes, M.D.E., Compton, S.G. and Cook, J.M. (1999) Sex J.M., Moore, J.C., Bossu-Dupriez, N., Chevolot, M. and allocation and local mate competition in Old World non- Michaloud, G. (2005) Clutch size: a major sex ratio deter- pollinating fig wasps. Behavioral Ecology and Sociobiology, minant in fig pollinating wasps? Comptes Rendus Biologie, 46, 95–102. 328, 471–476. Fisher, R.A. (1930) The Genetical Theory of Natural Selection. Maynard Smith, J. (1976) Evolution and the theory of games. Clarendon Press, Oxford, UK. 272 pp. American Scientist, 64, 41–45. Flanagan, K.E., West, S.A. and Godfray, H.C.J. (1998) Local Moore, J.C., Compton, S.G., Hatcher, M.J. and Dunn, A.M. mate competition, variable fecundity and information use in (2002) Quantitative tests of sex ratio models in a pollinating a parasitoid. Animal Behaviour, 56, 191–198. fig wasp. Animal Behaviour, 64, 23–32. Frank, S.A. (1985) Hierarchical selection theory and sex ratios. Moore, J.C. and Greeff, J.M. (2003) Resource defence in female II. On applying the theory, and a test with fig wasps. Evolution, pollinating fig wasps: two’s a contest, three’s a crowd. Animal 39, 949–964. Behaviour, 66, 1101–1107. Greeff, J.M. and Newman, D.V. (2011) Testing models of fac- Moore, J.C., Zavodna, M., Compton, S.G. and Gilmartin, P.M. ultative sex ratio adjustment in the pollinating fig wasp (2005) Sex ratio strategies and the evolution of cue use. Pro- Platyscapa awekei. Evolution, 65, 203–219. ceedings of the Royal Society B-Biological Sciences, 272, Hamilton, W.D. (1967) Extraordinary sex ratios. Science, 156, 1287–1294. 477–488. Patino, S., Herre, E.A. and Tyree, M.T. (1994) Physiological de- Hamilton, W.D. (1979). Wingless and fighting males in fig terminants of Ficus fruit temperature and implicaitons for wasps and other . Reproductive Competition, Mate survival of pollinator wasp species: Comparative physiol- Choice and Sexual Selection in Insects (eds. M.S. Blum & ogy through an energy budget approach. Oecologia, 100, N.A. Blum), pp. 167–220. Academic Press, New York, USA. 13–20. Hardy, I.C.W. and Hardy, I. (2002) Sex Ratios: Concepts and Pereira, R.A.S. and Prado, A.P.(2006) Effect of local mate com- Research Methods. Cambridge University Press, Cambridge, petition on fig wasp sex ratios. Brazilian Journal of Biology, UK. 424 pp. 66, 603–610. Herre, E.A, Jander, K. and Machado, C. (2008) Evolutionary Raja, S., Suleman, N., Compton, S. and Moore, J. (2008) The ecology of figs and their associates: recent progress and out- mechanism of sex ratio adjustment in a pollinating fig wasp. standing puzzles. Annual Review of Ecology and Systematics, Proceedings of the Royal Society B-Biological Sciences, 275, 39, 439–458. 1603–1610. Herre, E.A. (1985) Sex ratio adjustment in fig wasps. Science, Sheldon, B.C. and West, S.A. (2002) Sex ratios. Encyclopedia of 228, 896–898. Evolution (ed. M. Pagel), pp. 1040–1044. Oxford University Herre, E.A. (1987) Optimality, plasticity and selective regime in Press, Oxford, UK. figwaspsexratios.Nature, 329, 627–629. Somjee, U., Ablard, K., Crespi, B., Schaefer, P. and Gries, G. Hu, H.Y.,Jiang, Z.F.,Niu, L.M., Fu, Y.G.,Peng, Z.Q. and Huang, (2011) Local mate competition in the solitary parasitoid wasp D.W. (2009) Different stimuli reduce attraction to pollinators Ooencyrtus kuvanae. Behavioral Ecology and Sociobiology, in male and female figs in the dioecious fig, Ficus hispida. 65, 1071–1077. Biotropica, 41, 717–720. Steiner, S. and Ruther, J. (2009) How important is sex for fe- Hu, H.Y., Niu, L.M., Ma, G.C., Fu, Y.G., Peng, Z.Q. and Huang, males of a haplodiploid species under local mate competition? D.W.(2010) Permeability of receptive fig fruits and its effects Behavioral Ecology, 20, 570–574. on the re-emergence behaviour of pollinators. Ecological En- Sun, B.F., Wang, R.W. and Hu, Z. (2009) Ovipositing pattern tomology, 35, 115–125. of the fig wasps and its effect on the offspring sex ratio. Ihaka, R. and Gentleman, R. (1996) R: A language for data anal- Zoological Research, 30, 559–564. ysis and graphics. Journal of Computational and Graphical Werren, J.H. (1987) Labile sex ratios in wasps and bees. Bio- Statistics, 5, 299–314. Science, 37, 498–506. King, B.H. (1993) Sex ratio manipulation by parasitoid wasps. West, S.A. (2009) Sex Allocation. Princeton University Press, Evolution and Diversity of Sex Ratio in Insects and Mites Princeton, USA. 466 pp. (eds. D. Wrensch & M. Ebbert), pp. 418–441. Chapman and West, S.A., Herre, E.A. and Sheldon, B.C. (2000) The benefits Hall, New York, USA. of allocating sex. Science, 290, 288–290.

C 2012 The Authors Insect Science C 2012 Institute of Zoology, Chinese Academy of Sciences, 20, 228–234 234 H. Y. Hu et al.

West, S.A., Shuker, D.M. and Sheldon, B.C. (2005) Sex-ratio hispida in the tropical rainforests of Xishuangbanna, China. adjustment when relatives interact: a test of constrains on Acta Botanica Sinica, 44, 519–526. adaptation. Evolution, 59, 1211–1228. Yang, D.R., Peng, Y.Q., Song, Q.S., Zhang, G.M., Wang, R.W., Zhao, T.Z. and Wang, Q.Y.(2002) biology of Ficus Accepted November 9, 2011

C 2012 The Authors Insect Science C 2012 Institute of Zoology, Chinese Academy of Sciences, 20, 228–234